Discussion
The SNBTS currently provides allogeneic third-party EBV-specific T cells for patients with relapsed/refractory post-transplant lymphoproliferative disease (PTLD). As of November 2020 more than 100 patients with relapsed or refractory PTLD have been treated from the current bank under a Specials license, with a mean overall survival rate of over 40% at three years post treatment. Patients in this cohort with PTLD arising after solid organ transplant had better outcomes, with survival of over 60% at three years post-treatment, and with minimal adoptive cell therapy-related side-effects [5].
The current bank of EBV-specific T cells was manufactured from 2007-2014, and changes to GMP standards since this period have driven a requirement to optimize and refine the current manufacturing processes [20]. More recent methods for generation utilise cytokine release assays to capture virus-specific T cells, though this requires a different expansion process. LCL-based stimulation and expansion protocols are still in regular use for development of anti-cancer therapies [34-36], and therefore there is a need to identify optimal methods for production and analysis.
In this study we demonstrated that optimization of the standard autologous LCL-based method of EBV T cell manufacturing to a fully GMP-compliant closed-process process is feasible without compromise in quality of final cell product. The modifications to protocol, reagents and culture process were assessed principally using flow cytometry, which provides a rapid and quantitative method for analysis. Robust, validated flow cytometric assays are a cornerstone of effective reproducible cell therapy manufacture [21]. The use of flow cytometric analysis and functional profiling of EBV-specific T cells through cytokine expression in this study resulted in improved characterization of both start material and final product, and effective assessment of in-process culture optima, which has been used for analysis of other T cell therapeutics including a SARS-CoV-2 T cell product for COVID-19 treatment [37].
Intracellular cytokine staining for IL-2, TNF-α and IFN-γ provides a reliable method for discriminating the differentiation state of T cells [19, 22]. The combination of multi-parameter cytokine secretion-based phenotyping with t-SNE analysis forms a powerful tool for dissecting functional subpopulations within the CD8+ cytotoxic T cell compartment, and was used as the basis for analysis of improvements and refinements in manufacturing of the current SNBTS EBV-specific T cell therapy used for treatment of PTLD [20].
Using the combined surface marker and intracellular cytokine flow cytometric phenotyping approach we were able to identify that multiple rounds of LCL stimulation were unnecessary, and that extending stimulation may increase the level of anergy or loss of function in the T cells, as identified by a loss of absolute IFN-γ secretion and increased expression of CD57, a marker of terminal effector differentiation [23]. However, the reduction in stimulation round to maximize functional responses needs to be balanced with the requirement for high yields of cells for treatment of multiple patients from a single manufacturing run.
Adoptive T cell therapy relies on large-scale expansion of functional T cells to manufacture clinically relevant numbers for patient infusion, conventionally through use of standard culture flasks or gas-permeable bags. The introduction of large volume, high gas exchange culture vessels (G-Rex flask, Wilson Wolf) has significantly improved the rate and extent of T cell expansion capacity [24]. The G-Rex flasks are GMP-compliant and are scalable up to 1L flasks which are qualified as an FDA Class 1 medical device allowing full closed process manufacture. This closed process manufacture involves suitable sealed flasks, transfer bags, heat sealed tubing and the GatheRex cell harvester pump (Wilson Wolf) to ensure sterility in the clinical product. We identified that cell yields could also be improved by using G-Rex flasks for culture with no significant changes in phenotype. A minor change in T cell composition was identified in the G-Rex cultures, with increase in the percentage of CD8 cells. This consistency of final product phenotype was also retained when all reagents were converted to fully GMP-compliant standards. GMP-compliant medium and cytokines with no exogenous xenoproteins ensured that the modified process complied with current regulatory requirements. T cells generated with GMP compliant reagents and flasks suitable for closed process culture had a significant increase in retention of the TCM compartment. This has advantages for persistence of the cell therapy once administered to a patient [25,26].
A principal concern with the current LCL-based stimulation process is that high LCL (and therefore viral antigen load) ratio to T cells combined with multiple rounds could drive T cell exhaustion [27,28] and the reduced T cell: LCL ratio process outlined here quantified whether this resulted in functional differences. The LCL process appears robust, as reduced intensity stimulation over three rounds did not significantly affect the phenotype of the T cells at end-point, although the reduced ratio exhibited a significantly enhanced CD8+ cell secretion of IFN-γ and TNF-α. The only modulation of culture processes that was not undertaken was to replace or supplement IL-2 with other gamma-chain specific T cell growth factors. However, other studies have concluded that changing the cytokine-mediated expansion method from IL-2 to other cytokines such as IL-7, IL-15, or IL-21 has no significant effect on the overall phenotype or function of T cells for therapy [29]. The increased production of IFN-γ and TNF-α in response to stimulation in the reduced intensity LCL stimulation may suggest products made using this protocol could have increased effector functions against viral-infected cells following patient engraftment.
A key feature of this work was to identify a robust panel of surface and intracellular markers which could effectively classify the T cell differentiation status and development from initial material through to final product. Our approach supplies clear data for this, and demonstrates the utility of this approach for T cell therapies [37]. In addition, the use of t-SNE dimensionality reduction was very effective at condensing multiple parameters into a single image which could be used to identify the status of the material at any stage of manufacture. These images are both illustrative and quantitative and could therefore be used as part of a standardised product release process. This cytometric phenotype and analysis approach is sufficiently adaptable and inclusive that it would suitable for phenotypic and functional assay of other cell therapies including virus-specific and genetically-modified T cell therapies [30-33,37].